From mboxrd@z Thu Jan  1 00:00:00 1970
From: "Russ Cox" <rsc@plan9.bell-labs.com>
To: 9fans@cse.psu.edu
Subject: Re: Tr: [9fans] off topic: troff book
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Message-Id: <20010117152404.368F3199EF@mail.cse.psu.edu>
Date: Wed, 17 Jan 2001 10:24:01 -0500
Topicbox-Message-UUID: 4f15d7d8-eac9-11e9-9e20-41e7f4b1d025

This is getting more off topic by the day.  It is
not worth arguing categorically for or against
literate programming.  

At Bell Labs the programs I write are in the
traditional style Boyd is espousing.  
At Harvard I've used literate programming tools
to write a compiler, numerous small programs in
various languages, and an intro to programming
languages text demonstrating mark-and-sweep and
copying garbage collection in a from-scratch scheme
interpreter.

It is possible to write very clear code that really
doesn't need comments nor "literate" interpretation.
It is possible to use LP to great advantage for making
an unfamiliar programming style immediately accessible
to a new audience (like students in an intro course,
or a professor grading your assignments).
It is also possible to write completely inscrutable
code that literate programming tools alone have
no chance of saving.

The right approach for the job depends critically
on the intended audience and on the existence
of common coding style between you and that audience.
Most programmers cannot just dive into unfamiliar
code and poke around to figure out what's going on.
LP can make the dive a bit less scary.  Like any tool,
though, if used badly it can do more harm than good.

I'd be happy to discuss specifics off-9fans.

Russ